Global and Planetary Change 121 (2014) 78–88 Contents lists available at ScienceDirect Global and Planetary Change journal homepage: www.elsevier.com/locate/gloplacha History of land use in India during 1880–2010: Large-scale land transformations reconstructed from satellite data and historical archives Hanqin Tian a,⁎, Kamaljit Banger a,TaoBoa, Vinay K. Dadhwal b a International Center for Climate and Global Change Research, School of Forestry and Wildlife Sciences, Auburn University, Auburn, AL 36849, USA b National Remote Sensing Centre (NRSC), Indian Space Research Organisation, Balanagar, Hyderabad 500625, Andhra Pradesh, India article info abstract Article history: In India, human population has increased six-fold from 200 million to 1200 million that coupled with economic Received 27 January 2014 growth has resulted in significant land use and land cover (LULC) changes during 1880–2010. However, large dis- Received in revised form 2 July 2014 crepancies in the existing LULC datasets have hindered our efforts to better understand interactions among Accepted 9 July 2014 human activities, climate systems, and ecosystem in India. In this study, we incorporated high-resolution remote Available online 16 July 2014 sensing datasets from Resourcesat-1 and historical archives at district (N = 590) and state (N = 30) levels to – Keywords: generate LULC datasets at 5 arc minute resolution during 1880 2010 in India. Results have shown that a signif- land use and land cover icant loss of forests (from 89 million ha to 63 million ha) has occurred during the study period. Interestingly, the India deforestation rate was relatively greater under the British rule (1880–1950s) and early decades after indepen- remote sensing dence, and then decreased after the 1980s due to government policies to protect the forests. In contrast to forests, historical land archives cropland area has increased from 92 million ha to 140.1 million ha during 1880–2010. Greater cropland expan- cropland expansion sion has occurred during the 1950–1980s that coincided with the period of farm mechanization, electrification, deforestation and introduction of high yielding crop varieties as a result of government policies to achieve self-sufficiency in food production. The rate of urbanization was slower during 1880–1940 but significantly increased after the 1950s probably due to rapid increase in population and economic growth in India. Our study provides the most reliable estimations of historical LULC at regional scale in India. This is the first attempt to incorporate newly developed high-resolution remote sensing datasets and inventory archives to reconstruct the time series of LULC records for such a long period in India. The spatial and temporal information on LULC derived from this study could be used by ecosystem, hydrological, and climate modeling as well as by policy makers for assessing the impacts of LULC on regional climate, water resources, and biogeochemical cycles in terrestrial ecosystems. © 2014 The Authors. Published by Elsevier B.V. This is an open access article under the CC BY license (http://creativecommons.org/licenses/by/3.0/). 1. Introduction fold increase in population (200 million to 1200 million) coupled with economic growth (especially after the 1950s) that has resulted in Human activities have altered the Earth's environment by changing LULC transformations (Richards and Flint, 1994; DES, 2010). For exam- the land use and land cover (LULC) in the past several centuries (Liu ple, Richards and Flint (1994) have reported that total forest area de- et al., 2005a, 2005b; Hurtt et al., 2006; Liu and Tian, 2010). LULC changes creased from 100 million ha to 81 million ha while cropland area are major driving forces for biogeochemical cycles, climate change, and increased from 100 million ha to 120 ha during 1880–1950. The tempo- food production from regional to global scales (Houghton and Hackler, ral pattern of deforestation during 1880–2000 has a major control over 2003; Feddema et al., 2005; Jain and Yang, 2005; Tian et al., 2012a; temporal pattern of carbon emissions due to land use change (Chhabra Tao et al., 2013). Since 1850, LULC change alone has contributed to ap- and Dadhwal, 2004). Therefore, accurate LULC estimation is key for un- proximately 35% of anthropogenic carbon dioxide (CO2) emissions derstanding interactions among human activities, climate systems, and across the globe (Houghton et al., 2012). However, these environmental ecosystem as well as for the formulation of policies at national level changes occur at multiple spatial and temporal scales that may highly (Houghton and Hackler, 2003; Tian et al., 2003; Arora and Boer, 2010). differ among regions. In the 20th century, India has experienced a 6- In India, detailed LULC dataset collected from village level survey and aggregated at district level (N = 590) is available only for the period of 1950–2010 from the Department of Economics and Statistics (DES), ⁎ Corresponding author. Tel.: +1 334 844 1059; fax: +1 334 844 1084. Government of India (DES, 2010). In addition, Richards and Flint E-mail address: [email protected] (H. Tian). (1994) have compiled the historical LULC archives including croplands, http://dx.doi.org/10.1016/j.gloplacha.2014.07.005 0921-8181/© 2014 The Authors. Published by Elsevier B.V. This is an open access article under the CC BY license (http://creativecommons.org/licenses/by/3.0/). H. Tian et al. / Global and Planetary Change 121 (2014) 78–88 79 forests, grasslands/shrublands, and built-up areas at state level (N = vegetation ranges from tropical evergreen in the south to the alpine 30) during 1880–1980. However, there are certain limitations of the in- meadows in the north, and from the deserts in the west to the evergreen ventory LULC datasets. For example, LULC datasets in the tabular forms forests in the north-east of India (Joshi et al., 2006). are inadequate for the use in climate, hydrological and biogeochemical models that require LULC in the gridded format (Feddema et al., 2005; 2.1. Land use and land cover databases Liu et al., 2008; Tian et al., 2010). On the other hand, the remote sensing techniques make it possible to monitor contemporary LULC pattern at In this study, we focused on the five dominant LULC types including high spatial resolution but only cover a relatively shorter time period. cropland, forest, grasslands/shrublands, wastelands, and built-up or set- In India, several coarse resolution LULC dataset products such as moder- tlement areas. Cropland category is defined as the land cultivated for ate resolution imaging spectroradiometer (MODIS; Loveland and crops including single season, double or triple crops, shifting cultivation, Belward, 1997; Hansen and Reed, 2000), GlobCover developed from horticultural plantations, and orchards. The Food and Agricultural Orga- Envisat's Medium Resolution Imaging Spectrometer (MERIS; Arino nization of the United Nations (FAO) has also included temporary fallow et al., 2008), and GLC2000 based on SPOT4 satellite (Bartholome and lands into the Agricultural Area category (FAO, 2013). However, we did Belward, 2005) are available for the recent years. In addition, a regional not include fallow lands in cropland category since fallow lands have a LULC dataset based on the Advanced Wide Field Sensor of Resourcesat-1 significantly different influence on the biogeochemical and hydrological has been developed at a spatial resolution of 56-m by the National Re- cycles (Tian et al., 2003). Forest category includes the area evergreen mote Sensing Agency, India during 2004–2010 (NRSA, 2007). Linking and deciduous trees with N10% canopy cover as well as degraded forest remote sensing data (short time series but high spatial resolution) and types that has b10% of the canopy cover. This definition is similar to the inventory data (long time series but coarse spatial resolution in tabular forest cover definition used by the National Remote Sensing Center, format) is also a big challenge (Verburg et al., 2011). India (NRSA, 2007). The built-up or settlement area is defined as the Recently, Banger et al. (2013) reported that contemporary total land occupied by buildings, roads and railways. In the historical ar- cropland and forest area estimated at state level from inventory DES chives, it is difficult to differentiate the grasslands, grazing areas, and was better represented by LULC datasets developed from Resourcesat- shrublands. Therefore, we classified the term grasslands/shrublands as 1 than global scale remote sensing datasets. It is difficult to generate the areas occupied by grasslands and permanent pastures, meadows, and the historical LULC datasets using coarse resolution global remote sens- shrublands. Wastelands include the area that cannot be brought under cul- ing datasets that have large discrepancies with the inventory datasets in tivation such as area covered by mountains, deserts, and ice caps. India. Therefore, it is imperative to integrate contemporary remote In this study, we used inventory LULC datasets available at district sensing datasets from Resourcesat-1 with the historical tabular archives (N = 590) and state level (N = 30) from different sources along with to generate more reliable and useful LULC datasets, which cover longer the LULC datasets developed from remote sensing datasets available time periods in India. Previously, several global scale LULC datasets have from the Advanced Wide-Field Sensor (AWiFS) of Resourcesat-1 to con- been developed by combining remote sensing and inventory land use struct LULC at 5 arc minute resolution during 1880–2010 (Richards records at state level in India (Ramankutty and Foley, 1999; Klein and Flint, 1994; DES, 2010; Table 1). The LULC generated in this study Goldewijk et al., 2011). In this study, we made the first attempt to inte- are represented in fractional forms which consists percentages of five grate the high-resolution satellite (Resourcesat-1 at 56-m resolution) LULC types (cropland, forest, grasslands/shrublands, wastelands, and and existing inventory datasets at district and state levels to generate built-up) in each grid cell.